Technology assists when memory falters

ANN ARBOR—Software under development at the University eventually could power computerized caregivers—nurses’ aides of the future—to remind the elderly or people with brain trauma to perform tasks, and even help them navigate their surroundings.

The technology, called Autominder, would supplement rather than replace human caregivers, says Martha Pollack, electrical engineering and computer science professor, whose team is developing the software. But the growing shortage of health care providers, ballooning population of aging baby-boomers and increasingly longer life spans mean computers could be invaluable aides in caring for people with cognitive disorders, Pollack says.

“We’re always going to need human caregivers,” she said. “With the increased percentage of older adults, there won’t be enough adults to provide full-time care.”

The future of the aging population is such a concern that on April 6, Pollack will testify before the Senate Committee on Aging about the challenges of developing such technology, and about how increased government support for such research is critical to its success.

The Autominder software has been used in hand-held computers slightly bigger than a personal digital assistant and in a mobile robot called Pearl, or Nursebot. It works by giving instructions or other guidance to the patient, who operates the device by touching or tapping it when the task is complete.

Eventually, researchers hope to use motion sensors and contact sensors with the software and computers to warn an elderly user if, for instance, his water inadvertently was left running or a back door was left wide open.

“The aging population is going to need help in all different areas,” said James MacBain, director for research relations at the College of Engineering. “[Pollack’s] elder care technology is critical to allowing people to remain self supporting in an autonomous lifestyle longer.”

So-called minder technology exists now, but it basically is a “glorified alarm clock,” Pollack said. Autominder uses artificial intelligence technology tailored to each user to issue personalized reminders from data it interprets about what the person has done and is supposed to do.

Consider a hypothetical situation in which an elderly patient is reminded to take medication that must be ingested on a full stomach. Traditional methods simply would remind the patient to take the medicine at a specific clock time.

But by using sensor data, the computer could tell if a person has eaten and remind him or her to take the medication at the appropriate time, after the meal.

The University researchers are collaborating with a team from the University of Pittsburgh and Carnegie Mellon to develop the technology.

Pollack is a fellow of the American Association for Artificial Intelligence and editor-in-chief of the Journal of Artificial Intelligence Research.

Pollack and her research assistants are available. For more information on Pollack and her research, or the artificial, visit:http://ai.eecs.umich.edu/ http://www.eecs.umich.edu/~pollackm/ http://www.eecs.umich.edu/~pollackm/Pollack-web_files/nursebot/index.htm